Commits

Branch selector

master

User selector

All users

Datepicker

All time

Commit history

Commits on May 29, 2026

• download: add option to skip_download (#23059)

  Show description for 06d26df
  ngxson

  authored
  06d26dfView commit details

  Copy full SHA for 06d26df

  Browse repository at this point

• mtmd: Add DeepSeekOCR 2 Support (#20975)

  Show description for da3f990
  

  sfallah

  and

  ngxson

  authored
  da3f990View commit details

  Copy full SHA for da3f990

  Browse repository at this point

• CUDA: Check PTX version on host side to guard PDL dispatch (#23530)

  Show description for 6ed481e
  

  ORippler

  and

  JohannesGaessler

  authored
  6ed481eView commit details

  Copy full SHA for 6ed481e

  Browse repository at this point

• server: bump timeout to 3600s (#23842)

  Show description for cb47092
  ngxson

  authored
  cb47092View commit details

  Copy full SHA for cb47092

  Browse repository at this point

• model : support for DeepseekV32ForCausalLM with generic DeepSeek Sparse Attention (DSA) implementation (#23346)

  Show description for 1f0aa2a
  

  4 peopleauthored
  1f0aa2aView commit details

  Copy full SHA for 1f0aa2a

  Browse repository at this point

• llama: use f16 mask for FA to save VRAM (#23764)

  Show description for 031ddb2
  am17an

  authored
  031ddb2View commit details

  Copy full SHA for 031ddb2

  Browse repository at this point

• sync : ggml

  ggerganov

  committed
  fe12e42View commit details

  Copy full SHA for fe12e42

  Browse repository at this point

• ggml : bump version to 0.13.1 (ggml/1523)

  ggerganov

  committed
  ea02bc3View commit details

  Copy full SHA for ea02bc3

  Browse repository at this point

• ngram-mod : Add missing include (#23857)

  Show description for b000431
  oazizi000

  authored
  b000431View commit details

  Copy full SHA for b000431

  Browse repository at this point

• llama: add llm_graph_input_mtp (#23643)

  Show description for eef59a7
  

  am17an

  and

  ggerganov

  authored
  eef59a7View commit details

  Copy full SHA for eef59a7

  Browse repository at this point

• app : move licences to llama-app (#23824)

  Show description for 98e480a
  angt

  authored
  98e480aView commit details

  Copy full SHA for 98e480a

  Browse repository at this point

• cuda : disables launch_fattn PDL enrollment due to compiler bug (#23825)

  aendk

  authored
  241cbd4View commit details

  Copy full SHA for 241cbd4

  Browse repository at this point

• meta : Add missing `buffer` set in allreduce fallback !COMPUTE clear (#23480)

  Show description for 33c718d
  TheBlueMatt

  authored
  33c718dView commit details

  Copy full SHA for 33c718d

  Browse repository at this point

Commits on May 28, 2026

• hexagon: basic/generic op fusion support and RMS_NORM+MUL fusion (#23835)

  Show description for 19e92c3
  max-krasnyansky

  authored
  19e92c3View commit details

  Copy full SHA for 19e92c3

  Browse repository at this point

• mtmd-debug: add color and rainbow mode (#23829)

  Show description for 751ebd1
  ngxson

  authored
  751ebd1View commit details

  Copy full SHA for 751ebd1

  Browse repository at this point

• mtmd: fix gemma 4 projector pre_norm (#23822)

  ngxson

  authored
  c8914adView commit details

  Copy full SHA for c8914ad

  Browse repository at this point

• opencl: move backend info printing into its own function (#23702)

  Show description for 408ae2b
  lhez

  authored
  408ae2bView commit details

  Copy full SHA for 408ae2b

  Browse repository at this point

• ci : run ui publish on ubuntu-slim (#23818)

  Show description for 3ef2369
  CISC

  authored
  3ef2369View commit details

  Copy full SHA for 3ef2369

  Browse repository at this point

• ui: fix audio and video modality detection (#23756)

  Show description for 2f6c815
  ValdikSS

  authored
  2f6c815View commit details

  Copy full SHA for 2f6c815

  Browse repository at this point

• ci : releases use Github-hosted builds for the UI (#23823)

  Show description for 445b7ce
  ggerganov

  authored
  445b7ceView commit details

  Copy full SHA for 445b7ce

  Browse repository at this point

• app : improve help output (#23805)

  Show description for 479a9a1
  angt

  authored
  479a9a1View commit details

  Copy full SHA for 479a9a1

  Browse repository at this point

• mtmd: n_head_kv defaults to n_head (#23782)

  Show description for 0b56d28
  sfallah

  authored
  0b56d28View commit details

  Copy full SHA for 0b56d28

  Browse repository at this point

• mtmd: fix gemma 4 audio rms norm eps (#23815)

  Show description for d6be315
  

  ngxson

  and

  CISC

  authored
  d6be315View commit details

  Copy full SHA for d6be315

  Browse repository at this point

• ci : change Vulkan builds to Release to reduce ccache (#23820)

  Show description for dd15579
  ggerganov

  authored
  dd15579View commit details

  Copy full SHA for dd15579

  Browse repository at this point

• arg: Add LLAMA_ARG_API_KEY_FILE environment variable for --api-key-file (#23167)

  kucharskim

  authored
  7fb1e70View commit details

  Copy full SHA for 7fb1e70

  Browse repository at this point

• test-llama-archs: fix table format [no release] (#23810)

  JohannesGaessler

  authored
  d374e71View commit details

  Copy full SHA for d374e71

  Browse repository at this point

• ggml: auto apply iGPU flag CUDA/HIP if integrated device (#23007)

  fl0rianr

  authored
  30af6e2View commit details

  Copy full SHA for 30af6e2

  Browse repository at this point

• mmvq Optim: add MMVQ_PARAMETERS_TURING(mmvq_parameter_table_id) for … (#23729)

  Show description for d7be461
  

  3 peopleauthored
  d7be461View commit details

  Copy full SHA for d7be461

  Browse repository at this point

• CUDA: route batch>=4 quantized matmul to MMQ on AMD MFMA hardware (#23227)

  Show description for bc81d47
  

  jadenmach2

  and

  JohannesGaessler

  authored
  bc81d47View commit details

  Copy full SHA for bc81d47

  Browse repository at this point

• server: minor tweaks to use more cpp features (#23785)

  Show description for 0b24686
  mfuntowicz

  authored
  0b24686View commit details

  Copy full SHA for 0b24686

  Browse repository at this point

• hexagon: minor refresh for HMX FA and MM (#23796)

  Show description for a919001
  

  max-krasnyansky

  and

  trivikram-reddy1

  authored
  a919001View commit details

  Copy full SHA for a919001

  Browse repository at this point

• vulkan: fast path for walsh-hadamard transform (#23687)

  Show description for 48e7078
  jeffbolznv

  authored
  48e7078View commit details

  Copy full SHA for 48e7078

  Browse repository at this point

• chat : add Granite 4.1 chat template (#23518)

  jesus-talavera-ibm

  authored
  bb771cbView commit details

  Copy full SHA for bb771cb

  Browse repository at this point

• vulkan: fix wrong index variable in inner loop (#23665)

  winstonma

  authored
  7c48fb8View commit details

  Copy full SHA for 7c48fb8

  Browse repository at this point

• vulkan: Fix memory logger unsafe iterator access (#23667)

  winstonma

  authored
  91eb8f4View commit details

  Copy full SHA for 91eb8f4

  Browse repository at this point